ACRAdvanced Cipher Research
Software development

The model is half the work. We build the other half too.

Every statistical method we deliver ships inside software that ingests real data, runs reproducibly, and is evaluated continuously. We write that software ourselves, which is why it matches the mathematics.

Network graph illustration
What we build

Capabilities.

Data pipelines

Ingestion of multi-schema vendor feeds, columnar storage, partitioning by day and region, and the reconciliation logic that handles late, duplicated and contradictory records.

Routing and map-matching services

Self-hosted routing engines with national road tiles, map-matching of noisy traces, travel-time matrices and route-likelihood scoring exposed as internal services.

Evaluation harnesses

Reproducible experiment runners with frozen inputs, planted truth, versioned outputs and reports that a non-specialist can read. The harness is delivered with the model.

Privacy-aware processing

Aggregation that measures what individual information survives, strict separation of identifiers from attributes, and audit trails for every transformation.

High-throughput compute

Pipelines sized for tens of gigabytes a day and millions of records an hour, on commodity cloud or on-premise hardware, with cost tracked per run.

Deployment and operations

Infrastructure as code, signed releases, monitoring and alerting, and the runbooks to hand a system over to an in-house team.

Covariance contours
How we work

Small team, short loops, written decisions.

  • Proposal first. A two-page plain-English digest of what will be built, what it costs and what could go wrong, before any code.
  • Artefacts over status updates. Every milestone is a runnable thing with a report, not a slide.
  • Evidence before claims. "Done" means deployed and checked against real use, not code-complete.
  • Independent review. Models and code are challenged by a second set of eyes before they reach you.
  • Handover built in. Documentation, infrastructure code and runbooks are part of the deliverable from day one.
Stack

Tools we reach for.

Python and the scientific stackGo and Rust for servicesKotlin and Swift for mobilePostgreSQLParquet and DuckDBValhalla routingStan, PyMC and JAXLinux, containers, infrastructure as codeWireGuard and modern TLSCloudflare, Hetzner and Vultr

We are not attached to any of them. The data and the deployment constraints choose the tools.

Contact

Bring us a problem that resists the obvious model.

We take on a small number of engagements at a time. Tell us what you are measuring, what you need to decide, and what the data looks like today. We reply within two working days.

Email
info@advcipher.com

Office
IFZA, Dubai, United Arab Emirates
Working with clients in the UK, Europe and the Gulf.

What to include
The decision you need to support, the data sources you have (or want), timescales, and any constraints on privacy or deployment.

Start a conversation